This study is designed to document the relationship between serious mental illness and course of homelessness among adults, and to assess the contribution of various characteristics, behaviors, conditions, and events which may mediate the relationship. A three-wave panel design will be used to study exit patterns of homeless adults. This longitudinal study wt provide critical information for designing intervention strategies for homeless adults with mental illness. This study is designed (1) to document transition patterns (i.e., variations in rate of exit, duration of exit, and vulnerability to subsequent homelessness) for mentally ill and other homeless adults; (2) to document service utilization and Institutional contact to test their contributions to exit behavior; (3) to evaluate the degree to which sustained homelessness affects functional status which in turn may affect exit behavior; and (4) to examine the individual attributes and experiences that distinguish persons with chronic histories of homelessness. Severe mental illness is defined as schizophrenia or major affective disorder as defined by DSM-III-R criteria. Respondents will be interviewed three times over a period of eight months. A probability sample of 400 recently homeless adults (i.e. those whose current episode is less than 14 days) will be selected from service sites throughout Alameda County, California, selected because of its demographically diverse homeless population. Data on factors expected to mediate the relationship between mental health status and course of homelessness will be collected including alcohol and other drug use histories, treatment histories, human capital, material resources, personal attitudes, functional status and service utilization. Exit patterns for mentally ill and other homeless adults will be documented, as will variations by gender and racial or ethnic groups.